Image Classification on ImageNet 1k V2 (test)
75.84AccuracyLion
Evaluation results
| Method | Links | |
|---|---|---|
| LionModel=ViT-L/16, Resolution=384, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Pre-train on ImageNet-21K then fine-tune on ImageNet2023.02 | 75.84 | |
| AdamWModel=ViT-L/16, Resolution=384, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Pre-train on ImageNet-21K then fine-tune on ImageNet2023.02 | 75.1 | |
| LionModel=ViT-B/16, Resolution=384, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Pre-train on ImageNet-21K then fine-tune on ImageNet2023.02 | 74.06 | |
| AdamWModel=ViT-B/16, Resolution=384, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Pre-train on ImageNet-21K then fine-tune on ImageNet2023.02 | 73.81 | |
| LionModel=ViT-B/16, Optimizer=Lion, RandAug + Mixup=true, Training Protocol=Train from scratch on ImageNet2023.02 | 69.19 | |
| AdamWModel=ViT-B/16, Optimizer=AdamW, RandAug + Mixup=true, Training Protocol=Train from scratch on ImageNet2023.02 | 68.14 | |
| LionModel=ViT-S/16, Optimizer=Lion, RandAug + Mixup=true, Training Protocol=Train from scratch on ImageNet2023.02 | 67.68 | |
| AdamWModel=ViT-S/16, Optimizer=AdamW, RandAug + Mixup=true, Training Protocol=Train from scratch on ImageNet2023.02 | 66.73 | |
| LionModel=ViT-B/16, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 64.81 | |
| AdamWModel=ResNet-50, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 64.24 | |
| LionModel=ViT-S/16, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 64.14 | |
| LionModel=ResNet-50, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 64.02 | |
| SGDModel=ResNet-50, Optimizer=SGD, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 63.93 | |
| AdamWModel=ViT-S/16, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 63.09 | |
| AdamWModel=ViT-B/16, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 61.87 | |
| LionModel=Mixer-B/16, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 55.94 | |
| LionModel=Mixer-S/16, Optimizer=Lion, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 55.75 | |
| AdamWModel=Mixer-S/16, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 55.01 | |
| AdamWModel=Mixer-B/16, Optimizer=AdamW, RandAug + Mixup=false, Training Protocol=Train from scratch on ImageNet2023.02 | 53.37 |